Shape Feature Based Image Retrieval Research And Applications

With the development of Inerternt and multimedia technology products a large number of multimedia data which include images, audio, video and other information. How to retrieve the desired content quickly an efficiently from a large number of informantion has become a hotspot issue in the field of multimedia research. In order to solve the issue, people come up with Content-based Image Retrieval.Content-based image retrieval is the use of the nature of the image such as color, texture,shape, spatial relations, and image processing technology, pattern recongnition and computer vision technology etc to achieve image retrieval. Due to the shape feature of image is better meet people's visual perception and the users as well as prefer to retrieve image with it. However, the realization of shape-based retrieval has a certain degree of difficulty because of the intrisic specificity of shape feature. Therefore, Shape-based Image Retrieval, a challenging subject in CBIR, has very important significance of the research and future prospects.The paper firstly introduce the reserch background, status and the applied related technology based on the CBIR and Shape-based Image Retrieval, especially deeply investigate the crucial technology of Shape-based Image Retrieval.As the JSEG algorithm makes the image segmentation process complex and computationally expensive,we propose a novel method based on joint JSEG and watershed algorithm to overcome this drawback.Then for the troditional point-based image retrieval compute expensive and its accuracy is low, we propose two new image retreieval methods jonining the local point fetures and segmentation shape: image retreieval based on segmentation edge and SIFT feature point or image retreieval based on segmentation edge and SURF feature point. We extract the SIFT or SURF features on the image segmented by the improved JSEG algorithm. Then we divide the extracted image features into edge and non-edge features according to the rules.And we implement image retrieval by feature points matching.Finally, we designed and implemented a prototype system of image retrieval based on the methods proposed in the paper.Then we verified its efficiency and practicality by comparing with the traditional image retrieval based SIFT or SURF feature point.